Hi!! I had exactly the same error with Hugin 2019. I tried with enblend 3.2 and the error doesn't appears any more. Hope that works for you.
El miércoles, 27 de junio de 2018, 18:33:42 (UTC-5), Benjamin Schnieders escribió: > > Hi all, > > a while ago, I had this nasty enblend behavior that I could only fix by > using, well, multiblend. (See old "Enblend Oddity" thread.) > > Still have the same problem, when stitching, enblend complains at some > point that there's a too high risk for a defective seam line. > > Was unable to provide a minimal example this time, but uploaded a snapshot > while blending to > http://airesearch.de/files/staticimgs/enblendProb/ (500mb) > > After remapping, enblend loads and blends images 0001..0020 without > problem, then fails with 21. See the attached screenshot as well - all > images have sufficient overlap. > (I could convince it to also blend in 21 by re-arranging the order) > > > So next, it would blend 22 onto the rest, like so: > $ enblend -a -o result.tif --compression=DEFLATE --no-optimize > -f70994x8819+3868+645 step.tif medium0022.tif > enblend: info: combining non-overlapping images: step.tif 1/1 > enblend: info: combining non-overlapping images: medium0022.tif 1/1 > enblend: encountered degenerate image/mask geometry; too high risk of > defective seam line > enblend: info: remove invalid output image "result.tif" > $ enblend --show-signature > Version 4.2-1ubuntu1~xenial/xenial - Wed, 11 May 2016 22:27:29 +0200 > > > (If someone could confirm this behavior please?) > > Again, re-arranging the order "fixes" the problem, but I'd a) very much > prefer to blend all my images in one go, without trying all possible > combinations, and b) I want to save the blending masks for different > exposures, which will become difficult with the manual incremental blend. > Tried all combinations (no)-optimize, -a, coarse/fine mask, different > generation algorithms... > The problem does not exist in lower resolutions, so the 500mb original > .tif is the best I can do. > > Suggestions? > > Best, > Benjamin > > > > > Benjamin Schnieders wrote: > > Hi everyone, > > > > I ran into an interesting little problem stitching a project of mine. I > attached a minimal example - from the size of the pano you can imagine > there are a lot more input images, but the error shows with the first 3 > images alone. > > > > I'm using Hugin 2017.0.0.eac5e8cc546e and enblend 4.2. > > > > In short, enblend refuses to blend the 3 images, for varying reasons. > > Typically, I have "--levels=29" set by default, with that, it complains: > > BasicImage::BasicImage(Diff2D size): size.x and size.y must be >= 0 > > (I guess this is a case of an error caught in the wrong spot.... or at > least the message does not help...) > > > > Dropping the "levels", it hints that I might want to try --pre-assemble, > then fails with > > > > enblend: info: loading next image: testing0000.tif 1/1 > > enblend: info: loading next image: testing0001.tif 1/1 > > enblend: warning: images do not overlap; they will be combined without > blending > > enblend: info: loading next image: testing0002.tif 1/1 > > enblend: encountered degenerate image/mask geometry; too high risk of > defective seam line > > > > Adding --pre-assemble does not change the outcome. I am aware that I'm > blending a disconnected set, in the whole pano, these images are eventually > connected by other images - also, blending image 0 and 1, 0 and 2, and 1 > and 2 works, so I would assume 0,1,2 to work as well - but nope, that > fails. > > > > Any idea how I can get my panorama to stitch after all? > > > > Thanks! > > Benjamin > > > > -- A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ --- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/2d23f5e1-0bd2-4cd1-a81a-e6bbfcffa00f%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
